How does this improve AI performance?

By providing a clean, hierarchical, and logically organized context, it reduces noise and ambiguity, leading to higher accuracy and fewer hallucinations in AI outputs.

What is the primary purpose of the Context Engineering skill?

It is designed to autonomously extract unstructured data from multiple sources and transform it into structured formats that are easily interpreted and utilized by AI models.

What output formats does this skill support?

The framework primarily generates structured data in YAML, JSON, and Markdown with metadata (frontmatter) for easy integration into dev workflows.

Can I control how deep the agent crawls websites?

Yes, the skill includes a crawling configuration where you can set maximum depth, target domain patterns, and the maximum number of pages per domain.

Is this skill suitable for building RAG systems?

Absolutely. It is specifically designed to facilitate the 'data preparation' phase of RAG by turning messy source material into clean, searchable, and structured content.

Context Engineering Framework

Name: Context Engineering Framework
Author: ShunsukeHayashi

byShunsukeHayashi

•

Web Scraping & Data Collection

Extracts and transforms unstructured information from various sources into structured, AI-interpretable formats.

The Context Engineering Framework is a specialized skill designed to automate the process of gathering and structuring data for AI use. It enables Claude to ingest content from diverse sources—including URLs, local files, and raw text—and convert them into hierarchical, organized documents (YAML, JSON, or Markdown). By providing standardized templates for source analysis and expert profiling, this skill ensures that context provided to AI models is accurate, well-structured, and logically grouped, making it an essential tool for building knowledge bases, RAG systems, and comprehensive project documentation.

Key Features

01Hierarchical content structuring with logical grouping and summarization

02Autonomous multi-source extraction from URLs, local files, and raw text

03Standardized templates for source analysis and expert profile definition

04Automated persistence to AI-friendly formats with metadata frontmatter

05Configurable web crawling with depth control and domain filtering

067 GitHub stars

Use Cases

01Converting disparate web documentation into a structured Markdown knowledge base for RAG pipelines

02Defining and standardizing expert domain context to improve LLM consistency and accuracy

03Automating the analysis and summarization of complex web resources for project research

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add shunsukehayashi/miyabi context-eng

For use in Claude.ai and ChatGPT

Download Skill